A photo of David Stiff

5 Reasons Why The Meg Was Expected to Sink (But Vault’s AI Projected a Winner)

David Stiff on May 15th, 2018

After a $45 million opening weekend, The Meg is officially a surprise hit. A day before release the Hollywood Reporter released an article titled, 'Big-Budget 'The Meg' Heads for Tepid $20M-Plus U.S. Debut'. Traditional box office tracking one day out predicted a bomb. For a movie with a production budget of $150 million, that would be disastrous.

On the other hand, we were under no such illusions. At Vault our AI system predicted a $47 million opening weekend — way back in March 2018 — after just analyzing the movie trailer.

So how did they get it so wrong, just one day before release, while we were so right months in advance?

1. People don't say what they really mean

Traditional tracking, survey data and social media monitoring all have inherent biases. They rely on people being truthful and unwavering in their answers. But people aren't. They’re fickle, for good or for bad. In the lead up to the 2016 Presidential elections, Donald Trump was trailing in the polls. Brexit wasn't expected to win either, but we all know what happened yet most of us don't understand why. The reason the polls were wrong was that there is an inherent bias in testing a small sample of the population and it creates a forecast that can be based on uncorrelated and even corrupted data.

And while The Meg isn't likely to win best screenplay, a lot of people wanted to see it, even if they might be embarrassed to tell their friends or coworkers.

2. Samples don't necessarily reflect a market

Unlike traditional predictive methods that rely on a sample-based approach and scale the results, artificial intelligence aims to analyze an entire market to deliver a forecast that is a true reflection of the audience.

Vault's AI platform is trained on 30 years worth of audience data from tens of thousands of movies. By analyzing The Meg trailer, Vault's AI platform was able to simulate how audiences would react to The Meg, how'd they influence others and ultimately how many tickets would be sold over the opening weekend. Simply put, Vault concluded that The Meg was a story that would sell tickets. And so, Vault predicted a strong $47 million opening weekend.

3. Traditional tracking methods haven't evolved with the movie business

Traditional audience tracking struggles to adapt its model for evolving or changing market conditions. The entertainment industry continues to grow and rapidly change each quarter.

Traditional tracking methods are slow to react to market changes — if at all.

They haven't been able to account for Warner Brothers releasing its movies to far greater audience sizes than before. Nor can they take into account that a Warner Brothers release looks very different to a Universal release. Just like a Disney release looks very different to a Lionsgate release. The reality of the media and entertainment landscape is that it continues to change and grow. Vault's AI system continuously adapts its movie audience model to the ever-changing factors that can impact movie attendance and revenue. Each and every minute Vault's AI system learns how the market is evolving and adjusts its model accordingly. And when it gets things wrong, it learns why it was wrong and then adjusts.

4. Uncorrelated Data

Studio executives and their marketing teams have an endless number of data sources and services that help them visualize an audience. But unfortunately, not all data is equally reliable. For many studio executives, it's very difficult to analyze how well a data source correlates to a particular movie audience.

So, why is correlated data important? Imagine if Google Maps were to tell you that it's only going to take you 20 minutes to drive home based on traffic data from another state? You’d be quite angry when the drive was not as expected and wasted your time and money. You’d probably stop using Google Maps right away.

While AI does make inaccurate predictions — and Vault's AI is no exception — supplying a consistent predictive service correlated to box office results is critical for optimizing the studio executive's decision-making processes.

5. In the end, it's all about the story

The three most important factors in the success of any movie are: the story, the story, and the story. Not even lead actor Jason Statham bemoaning the final cut uttering, "Where's the f***ing blood?" could have derailed The Meg.

And why was that? Because the majority of the audience was driven by purely story. They were going to The Meg because the story elements activated them. Story elements are the brand promise and messaging of the movie. It sets the expectations and frames the experience before it happens. For The Meg, all the elements of the story came together like a perfect storm — and it delivered exactly what audiences wanted, driving more and more people to the box office.

As "corny" as the story was supposed to be, traditional sample forecast methods couldn't understand that the story of The Meg actually had a massive audience waiting to see it. But AI could.